01 / HARDWARE BUYBACK

We buy GPUs & servers.

Upgrading your hardware or retiring a server?
Tell us what you have available.

01 / GPU

GPUs

Share the exact model, VRAM, quantity, and condition of the GPUs you would like to sell.

02 / SERVER

Servers

Share your server configuration, including CPUs, GPUs, memory, storage, and quantity.

Send the details. Discuss a buyback.

Include your location, photos or a specification list, and your asking price if available. Eligibility, valuation, and transaction details are confirmed after review.

Explore hardware buyback

NEOTERIC LAB / COMPUTECOMING SOON

FLAGSHIP / GPU INSTANCE

GB300MIG Instances

Start with a GPU partition.
Explore Grace Blackwell on your terms.

Indicative hourly rate

2.00 / MIG instance / h

Discuss a GB300 workload

In preparation. Rentals are not open yet.

DGX STATIONGB300 / 01

GRACE BLACKWELL ULTRAProduct illustration · Exxact

RENTAL UNITOne MIG instance

An isolated GPU partition on a DGX Station.

PARTITIONINGUp to 7 instances

Actual profiles and resource quotas to be announced.

WORKLOADSDevelopment & inference

Confirm GPU memory needs before choosing an instance.

02 / DEDICATED COMPUTE

Other ways to run.

A whole Mac, or a dedicated GPU.
Each price has a different rental unit.

APPLE SILICONCOMING SOON

Mac Studio

Indicative hourly rate

2.50 / machine / h

For local model inference and Apple Silicon development.

Memory: 256GB unified
Rental unit: Whole machine
To be confirmed: Chip generation · storage

Discuss this configuration Reference image

NVIDIA RTX PROCOMING SOON

RTX PRO 6000

Indicative hourly rate

1.95 / GPU / h

For CUDA inference, model development, and GPU experiments.

Server configuration: 2 GPUs
Rental unit: One GPU
To be confirmed: GPU edition · VRAM per GPU

Discuss this configuration Reference image

All amounts are reference prices; currency is awaiting confirmation. Billing rules and final quotes will be published before launch. Images illustrate product families, not confirmed delivered configurations.

03 / BEFORE YOU RENT

Know what you are getting.

These details are not yet finalized.
They will be published before rentals open.

Resources

MIG profile, GPU memory, CPU cores, RAM, and storage per instance.

Access & network

Region, remote access method, bandwidth, and traffic limits.

Billing & service

Minimum rental time, start/stop billing, data retention, and support.

MODEL API / COMING SOON

Model APIs are also planned.

Model lineup, context limits, and token prices will be announced after deployment validation.

Share your API needs